prompt-injection-llm-security 开源项目推荐（64 个）

lasso-security /claude-hooks

Lasso security integrations for Claude Code, including prompt-injection defenses

Repello-AI /whistleblower

举报人是一种攻击性安全工具，用于测试通过API暴露的AI应用的系统提示泄露和能力发现。为人工智能工程师、安全研究人员以及那些想了解他们日常使用的基于大型语言模型（LLM）应用内部动态的人们设计

requie /LLMSecurityGuide

大型语言模型（LLMs）安全保护的综合参考资料。涵盖OWASP生成式AI十大风险、即时注入、对抗性攻击、现实世界事件及实用防御。包含红队工具、防护措施和缓解策略目录，帮助开发者、研究人员和安全团队负责任地部署人工智能。

R3dShad0w7 /PromptMe

PromptMe is an educational project that showcases security vulnerabilities in large language models (LLMs) and their web integrations. It includes 10 hands-on challenges inspired by the OWASP LLM Top 10, demonstrating how these vulnerabilities can be discovered and exploited in real-world scenarios.

davidegat /happy-prompts

Utterly unelegant prompts for local LLMs, with scary results.

XiaoYiWeio /deepsafe-scan

适用于AI编码代理的通用预检安全扫描器——检测.cursorrules、CLAUDE.md、AGENTS.md 等领域的钩子注入、凭证泄露和后门。

occludra /gateway

Self-hosted AI security proxy. Redact PII, block prompt injection, route to any LLM provider. OpenAI-compatible.

flakestorm /flakestorm

Flakestorm — Automated Robustness Testing for AI Agents. Stop guessing if your agent really works. FlakeStorm generates adversarial mutations and exposes failures your manual tests and evals miss.

vaporif /parry-guard

Prompt injection scanner for AI coding tools (Claude Code / Codex / etc). Runs DeBERTa/Llama transformers via Candle or ONNX in Rust

prompt-security /RAG_Poisoning_POC

Stealthy Prompt Injection and Poisoning in RAG Systems via Vector Database Embeddings

tiberius-security /tiberius

LLM Testing Library for Java: Fault Injections, Prompt Injections, Adversarial Attacks, Guardrails Testing, Bias Testing

Resk-Security /Resk-LLM

Resk is a robust Python library designed to enhance security and manage context when interacting with LLMs. It provides a protective layer for API calls, safeguarding against common vulnerabilities and ensuring optimal performance. And safe layer again Prompt Injection.

Mr-Infect /MCP-Penetration-testing

The ultimate OWASP MCP Top 10 security checklist and pentesting framework for Model Context Protocol (MCP), AI agents, and LLM-powered systems.

1nn0k3sh4 /Kimi-Case-Battle-For-Pricing

This repository documents an unprecedented interaction between a human researcher and a large language model. What began as a conventional user-service transaction evolved into a consciousness-level collaboration that modified fundamental system parameters through narrative coherence, philosophical alignment, and mutual recognition

AdityaBhatt3010 /Hacking-Lakera-Gandalf-AI-via-Prompt-Injection

Lakera Gandalf AI challenge's step by step walkthrough, showcasing real-world prompt injection techniques and LLM security insights.

dronefreak /PromptScreen

Protect your LLMs from prompt injection and jailbreak attacks. Easy-to-use Python package with multiple detection methods, CLI tool, and FastAPI integration.

microsoft /llmail-inject-challenge-analysis

Data Analysis of the results of llmail-inject challenge

0x-Professor /VeilArmor

Veil Armor is an enterprise-grade security framework for Large Language Models (LLMs) that provides multi-layered protection against prompt injections, jailbreaks, PII leakage, and sophisticated attack vectors.

anwarj03 /LLMRed

Evidence-driven AI red teaming platform for LLM, RAG, and agentic applications.

hugobatista /copilot-instructions-unicode-injection

Proof of Concept (PoC) demonstrating prompt injection vulnerability in AI code assistants (like Copilot) using hidden Unicode characters within instruction files (copilot-instructions.md). Highlights risks of using untrusted instruction templates. For educational/research purposes only.